They're using a pretty fuzzy definition of "core", but otherwise this is a great post. The only thing lacking is comparison to bare metal servers with similar specs (as was mention in the EC2 gripefest the other day).

Author here, "core" is a fuzzy term I suppose, but it's often the most disclosed by services to the outside world. I agree it would be interesting to compare these metrics to some baremetal numbers. It would then be relatively easy to determine if they are dedicated cores of the type advertised.